Twenty years have passed since Freeman and Medoff's What Do Unions Do? This essay assesses their analysis of how unions in the U.S. private sector affect economic performance - productivity, profitability, investment, and growth. Freeman and Medoff are clearly correct that union productivity effects vary substantially across workplaces. Their conclusion that union effects are on average positive and substantial cannot be sustained, subsequent evidence suggesting an average union productivity effect near zero. Their speculation that productivity effects are larger in more competitive environments appears to hold up, although more evidence is needed. Subsequent literature continues to find unions associated with lower profitability, as noted by Freeman and Medoff. Unions are found to tax returns stemming from market power, but industry concentration is not the source of such returns. Rather, unions capture firm quasi-rents arising from long-lived tangible and intangible capital and from firm-specific advantages. Lower profits and the union tax on asset returns leads to reduced investment and, subsequently, lower employment and productivity growth. There is little evidence that unionization leads to higher rates of business failure. Given the decline in U.S. private sector unionism, I explore avenues through which individual and collective voice might be enhanced, focusing on labor law and workplace governance defaults. Substantial enhancement of voice requires change in the nonunion sector and employer as well as worker initiatives. It is unclear whether labor unions would be revitalized or further marginalized by such an evolution.